Project Summary Studies examining the relationship between psychiatric illness and suicide from the Department of Veterans Affairs indicate that individuals with bipolar disorder have the highest rate of suicide, which is even greater compared to post-traumatic stress disorder, schizophrenia and major depression. When left untreated, bipolar disorder has a 5-15% risk of death by suicide, making it a top priority for Veterans' mental health care. Individuals with bipolar disorder behave impulsively ? even during euthymic periods and understanding the relationship between impulsivity and suicidal behavior in bipolar disorder may assist in the identification of individuals at greatest risk for a future suicide attempt. Progress in understanding this relationship has been significantly hampered, however, by the conflation of different measures of impulsivity. Impulsivity can be measured as a relatively stable trait using self-report measures or as a state measure assessed through the use of behavioral paradigms sensitive to changing environmental contingencies. Models of state impulsivity have focused on (1) ?rapid-response inhibition? involving difficulty inhibiting responses that are prepotent in the context of changing environmental situations and (2) ?choice impulsivity? defined as the inability to delay gratification for a larger reward. These models correlate weakly, but strongly overlap with brain regions implicated in the neurobiology of bipolar disorder. We propose investigating measures of state and trait impulsivity in 40 euthymic Veterans with bipolar disorder (BD) and a suicide attempt history (BD+S), 40 euthymic Veterans with bipolar disorder and no suicide attempt history (BD-S) and 40 healthy controls (HC). The primary suicide outcome measure in this study will be suicide attempts assessed using the Columbia Suicide Severity Rating Scale and medical record review. This will be complemented through the use of a broader composite suicide outcome measure based on the occurrence of any of 5 types of suicidal behavior including: (1) death by suicide; (2) suicide attempt; (3) interrupted suicide attempt; (4) aborted suicide attempt; or (5) preparatory suicidal acts. Following comprehensive baseline clinical, neuropsychological and neuroimaging assessments all Veterans will subsequently complete 6- and 12-month in-person follow-up clinical assessments of mood and suicidal behavior. Novel measures of trait impulsivity in this study will include urgency and impulsive/premeditated aggression. We will assess the relationship between impulsive aggression and social cognition, functional disability and neurocognitive functioning in BD+S. We will further investigate two neural circuits tapping state measures of rapid response inhibition (using a go/no-go task) and choice impulsivity (using a delay discounting task), respectively, using functional magnetic resonance imaging and neurite orientation dispersion and density imaging, which will provide novel measures of fiber arrangement and neurite density in tracts connecting gray matter regions. Dynamic causal modeling will be used to empirically test causal models regarding the interaction of brain regions comprising these two circuits respectively. Machine learning will be used to integrate baseline trait and state measures of impulsivity to longitudinally predict suicidal behavior over one year. The specific aims of this study are: (1) to investigate the relationship between trait measures of impulsivity and suicide attempt history in Veterans with bipolar disorder; (2) to investigate the neural circuitry underlying two models of state impulsivity and their relationship to suicide attempt history in Veterans with bipolar disorder; and (3) to identify which combination of impulsivity measures differentiates BD+S from BD-S and HC at a baseline timepoint and can be used to predict suicidal behavior longitudinally over 1 year.